Hi, I am looking for a general advice to parallelize part of a code that I am writting. I have almost null experience on this matter, but I think that my case should be almost trivial.
I have a function F1 that takes several arguments (two objects of a custom TYPE and some numbers), and perform a mathematical computation which only depends of the paremeters, returning a DVec3. It is pretty fast, around 1 to 10 millons of executions per second (depending on a precision parameter) on my standard PC. I want to define a function F2, that takes only two objects as the those mentioned above, and makes around 10 calls to F1 (passing the two objects and specific previulsy known values to to the remaining parameters). Then, collect them in sequence of DVec3. Although a call to F2 is fast, I will need to call it millons of times (but not simultaneously). My wish is that every time I call F2, it computes in parallel the 10 calls of F1 (I mean, to excecute simultaneusly the F1s, each excecution in a different physical core). I need it works in both, C and JS backends. I prefere simplicity and compatiblity/maintainability over maximum efficiency. What is the way to go? I found may alternatives but I am not sure which should I choose (malebolgia, weave, taskpools, theads, threadpool). Thanks!
